Exploration of trait dataset

The objective of this document is to: - visually explore the trait database - understand trait coverage - identify data gaps - check for possible inconsistencies

Description of trait databases

# summary of NAs
num_na <- apply(is.na(traits), 2, sum)

db_na <- num_na[grep("original", names(num_na))]
names(db_na) <- gsub("original_taxa_", "", names(db_na))
# database with best match
par(mar = c(4, 8, 1, 1))
barplot(
  sort(db_na),
  horiz = TRUE,
  las = 1,
  xlab = "Number of missing species"
)

# summary of NAs
real_traits <- !grepl("original", names(traits)) &
  names(traits) != "accepted_taxa"

trait_na <- apply(is.na(traits[, real_traits]), 2, sum)

# database with best match
par(mar = c(4, 8, 1, 1))
barplot(
  sort(trait_na),
  horiz = TRUE,
  las = 1,
  xlab = "Number of missing species"
)

# species with lowest completness
sp_na <- apply(is.na(traits[, real_traits]), 1, sum)
names(sp_na) <- traits$accepted_taxa

boxplot(sp_na, ylab = "Number of missing traits")

hist(sp_na, xlab = "Number of missing traits", ylab = "Number of species")

mat_na <- !is.na(traits[, real_traits])
mat_na <- apply(mat_na, 2, as.numeric)
row.names(mat_na) <- traits$accepted_taxa
# heatmap(mat_na, scale = "none")

mat_na <- mat_na[order(sp_na), order(trait_na)]
plot_ly(
  x = colnames(mat_na),
  y = rownames(mat_na),
  z = mat_na,
  type = "heatmap"
) %>%
  layout(margin = list(l = 120))

Comparison

SLA

There are three sources of information for Specific leaf area (SLA) : Hodgson et al. 2023 (in mm2/mg), GIFT (in cm2/g) and BIEN (in m2/kg = mm2/mg).

# names(traits)[grep("SLA", names(traits))]
sla <- data.frame(
  "Hodgson" = traits$SLA.mm2.mg.1_Hodgson_2023,
  "GIFT" = traits$SLA.mean.cm2.g.1_GIFT / 10,
  "BIEN" = traits$leaf.area.per.leaf.dry.mass.m2.kg.1_BIEN
)

pairs(sla, lower.panel = panel.smooth, upper.panel = panel.cor)

Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter

Plant height

There are four sources of information for plant height.

# names(traits)[grep("Height", names(traits))]
# all in mm2/mg

height <- data.frame(
  "spvignes" = traits$Plant_Height_SPVignes,
  "lososova" = traits$Plant.height_m_Lososova_2023,
  "BIEN" = traits$whole.plant.height.m_BIEN,
  "GIFT" = traits$Plant.height.mean.m_GIFT
)

pairs(height, lower.panel = panel.smooth, upper.panel = panel.cor)

Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter
Warning in par(usr): argument 1 does not name a graphical parameter